search

amp-cli / amp

Control script for *amp-style stacks
6 stars 15 forks source link

readme

Download

sudo curl -LsS https://download.civicrm.org/amp/amp.phar -o /usr/local/bin/amp
sudo chmod +x /usr/local/bin/amp

Build and Test

The build and test processes are based on composer, phpunit, and box.

To facilitate local development and testing with multiple versions of php and mysqld (on Linux and macOS), the repo includes a few helpers: default.nix, ./scripts/run-tests.sh, and ./scripts/run-build.sh. These helper scripts require the nix package manager. Usage:

## Get the code
git clone https://github.com/amp-cli/amp
cd amp

## Start a shell with php+composer
nix-shell

## Run the test suites
./scripts/run-tests.sh

## Run the build, creating "dist/amp.phar"
./scripts/run-build.sh

About amp: Vision

amp is a tool to facilitate development of PHP web applications. The goal is to complement composer (and similar tools) by adding a (mostly) automated step to setup the database and webserver for newly downloaded code. For example, a developer checking out a project might say:

me@localhost:~/src$ composer create-project example/my-application --dev
me@localhost:~/src$ cd my-application
me@localhost:~/src/my-application$ ./bin/amp create --url=http://localhost:8003
URL: http://localhost:8003
Admin User: admin
Admin Password zFWx9D22

The my-application package depends on the amp package (using require-dev or suggest). The amp create step creates a new database in the local mysqld and a new virutal-host in the local httpd; then it writes out necessary credentials (eg the mysql username and password) to a config file.

Additional thoughts:

About amp: Pre-Alpha Example

At time of writing, amp is in-development and doesn't fully meet its vision. In the third line, the developer shouldn't call amp create directly; rather, the author of my-application should include an install.sh script, and the downstream developer can run it:

me@localhost:~/src$ composer create-project example/my-application --dev
me@localhost:~/src$ cd my-application
me@localhost:~/src/my-application$ ./bin/amp config
me@localhost:~/src/my-application$ ./bin/install.sh
Login to the application:
 * URL: ${AMP_URL}
 * Username: admin
 * Password: default

The amp config command determines how to connect to MySQL and httpd. It may scan the local system for common configurations (Ubuntu vs MAMP vs MacPorts; Apache vs nginx), prompt the user for information, and retain the info (in ~/.amp) for future use.

The install.sh is mostly specific to the application, but it builds on amp to address the tedious bit about setting up mysqld and httpd. For example, one might say:

#!/bin/bash
set -e
APPDIR=`pwd`

## Create a new database and virtual-host
eval $(amp create --root="$APPDIR/web")
amp datadir "$APPDIR/log" "$APPDIR/cache"

## Load DB
cat $APPDIR/sql/install.sql | mysql -u$AMP_DB_USER -p$AMP_DB_PASS $AMP_DB_NAME

## Create config file
cat > $APPDIR/conf/my-application.ini <<MYCONFIG
[mysql]
username=${AMP_DB_USER}
password=${AMP_DB_PASS}
database=${AMP_DB_NAME}
hostname=${AMP_DB_HOST}
MYCONFIG

echo "Login to the application:"
echo " * URL: ${AMP_URL}"
echo " * Username: admin"
echo " * Password: default"

Backlog

See doc/backlog.md

FAQ

Q: Is amp stable? Should I rely on it right now?

A: Probably not. amp is pre-alpha. Interfaces and workflows are likely to change.

Q: How do I configure amp to work on my system?

A: Run `amp config

Q: How do I know if amp is working?

A: Run amp test

Q: How does amp assign a virtual hostname and port?

A: You can specify one by passing the --url option to create. If omitted, it will use localhost and assign an alternative port.

Q: How does amp name databases and database users?

A: The name is computed by taking the directory name (eg my-application) and appending some random characters. The directory name may be truncated to meet MySQL's size limits. The name is the same for the DB and user.

Q: Where does amp store its configuration data?

A: By default, ~/.amp. If you define the environment variable AMPHOME, it will store in the specified directory.

Q: I have five web apps installed. How does AMP distinguish them?

A: Each application should have its own directory (eg /home/me/src/my-application-1). By default, amp assumes that each directory corresponds to a single virtual-host and a single MySQL database. If you need an additional virtual-host and DB for that application, call create again with the --name argument. If you want an additional virtual-host XOR DB, specify --skip-db or --skip-url.

Q: How do I build a stand-alone PHAR executable for amp?

A: Install Box. Then, in the amp source dir, run "php -d phar.readonly=0 which box build"

Internal Architecture

amp uses components from Symfony 2 (eg Console, Config, and Dependency-Injection).

There are a few key services defined in the container:

There may be competing implementations of db, httpd, hosts, and perm -- eg one implementation might connect to a remote mysqld while another launches a local mysqld on a ramdisk. These can be chosen at runtime by calling commands like:

## Set options interactively
amp config

## Set options individually
amp config:set --httpd_type=XXX
amp config:set --db_type=XXX
amp config:set --perm_type=XXX

## Set options en masse
amp config:set --httpd_type=XXX --db_type=XXX --perm_type=XXX

Parameters and services may be configured in amp's source-tree (app/defaults/services.yml) or in the local home directory (~/.amp/services.yml). Parameters entered through the CLI (amp config, amp config:set, etc) are stored in the local home directory (~/.amp/services.yml).